Difference between revisions of "Random name generation"

From RogueBasin
Jump to navigation Jump to search
((pos*4/len) would range from 1-3, giving 20-23% chance to quit, defeating the goal of that line to bias to short, memorable, names)
(Markov-Chains Approach in Python)
Line 347: Line 347:


As you can see, some are ok but most aren't. This is because the dataset although big isn't quite big enough. To compensate for this, you could try to implement artificial rules to make the names more readable, but I will leave that as an exercise for the reader.
As you can see, some are ok but most aren't. This is because the dataset although big isn't quite big enough. To compensate for this, you could try to implement artificial rules to make the names more readable, but I will leave that as an exercise for the reader.
==Markov-Chains Approach in Python==
Here is the piece of code used in [[Mines of Elderlore]] to generate random names:
import random
# from http://www.geocities.com/anvrill/names/cc_goth.html
PLACES = ['Adara', 'Adena', 'Adrianne', 'Alarice', 'Alvita', 'Amara', 'Ambika', 'Antonia', 'Araceli', 'Balandria', 'Basha',
'Beryl', 'Bryn', 'Callia', 'Caryssa', 'Cassandra', 'Casondrah', 'Chatha', 'Ciara', 'Cynara', 'Cytheria', 'Dabria', 'Darcei',
'Deandra', 'Deirdre', 'Delores', 'Desdomna', 'Devi', 'Dominique', 'Drucilla', 'Duvessa', 'Ebony', 'Fantine', 'Fuscienne',
'Gabi', 'Gallia', 'Hanna', 'Hedda', 'Jerica', 'Jetta', 'Joby', 'Kacila', 'Kagami', 'Kala', 'Kallie', 'Keelia', 'Kerry',
'Kerry-Ann', 'Kimberly', 'Killian', 'Kory', 'Lilith', 'Lucretia', 'Lysha', 'Mercedes', 'Mia', 'Maura', 'Perdita', 'Quella',
'Riona', 'Safiya', 'Salina', 'Severin', 'Sidonia', 'Sirena', 'Solita', 'Tempest', 'Thea', 'Treva', 'Trista', 'Vala', 'Winta']
###############################################################################
# Markov Name model
# A random name generator, by Peter Corbett
# http://www.pick.ucam.org/~ptc24/mchain.html
# This script is hereby entered into the public domain
###############################################################################
class Mdict:
    def __init__(self):
        self.d = {}
    def __getitem__(self, key):
        if self.d.has_key(key):
            return self.d[key]
        else:
            raise KeyError(key)
    def add_key(self, prefix, suffix):
        if self.d.has_key(prefix):
            self.d[prefix].append(suffix)
        else:
            self.d[prefix] = [suffix]
    def get_suffix(self,prefix):
        l = self[prefix]
        return random.choice(l) 
class MName:
    """
    A name from a Markov chain
    """
    def __init__(self, chainlen = 2):
        """
        Building the dictionary
        """
        if chainlen > 10 or chainlen < 1:
            print "Chain length must be between 1 and 10, inclusive"
            sys.exit(0)
   
        self.mcd = Mdict()
        oldnames = []
        self.chainlen = chainlen
   
        for l in PLACES:
            l = l.strip()
            oldnames.append(l)
            s = " " * chainlen + l
            for n in range(0,len(l)):
                self.mcd.add_key(s[n:n+chainlen], s[n+chainlen])
            self.mcd.add_key(s[len(l):len(l)+chainlen], "\n")
   
    def New(self):
        """
        New name from the Markov chain
        """
        prefix = " " * self.chainlen
        name = ""
        suffix = ""
        while True:
            suffix = self.mcd.get_suffix(prefix)
            if suffix == "\n" or len(name) > 9:
                break
            else:
                name = name + suffix
                prefix = prefix[1:] + suffix
        return name.capitalize() 
for i in range(100):
    print MName().New()


==Finite State Approach==
==Finite State Approach==

Revision as of 09:31, 27 August 2008

There are a lot of different ways to tackle random name generation. Some use existing lists of names to generate variants of these names, such as the Markov-Chain Approach. Other variants use lists of syllables to randomly select names. And others combine different words or parts of words into new random names.

Markov-Chains Approach

Markov-Chains are usually used in the field of thermodynamics, but I will be using a fudged version of Markov-Chains to generate some random names. The basic premise of Markov-Chains is that if we have a big enough data set we should be able to predict the next event in a series of seemingly random events. So using this we need to create a dataset of names that we want to generate, lucky for you I have already done this. Here is the List of Names.

This list is every male and gender-neutral name from babynames.com. Now we need to create a program to parse this list for it's statistics. To make this interesting I have split the statistics up, 1. Letter combinations that begin names, 2. Letter combinations that end names, and 3. Letter combinations that are in between the beginning and end. This should give us slightly better names. The following is my code samples in C++, after compilation this program will output some randomly generated names.


// CWordFrequency.h
#ifndef CWORDFREQUENCY_H_
#define CWORDFREQUENCY_H_

class CWordFrequency
{

   private:
       int countBeginning;
       int countEnd;
       int countWithin;

   public:
       CWordFrequency();
       ~CWordFrequency();
       void incrementCountBeginning();
       void incrementCountEnd();
       void incrementCountWithin();
       int returnCountBeginning();
       int returnCountEnd();
       int returnCountWithin();

};

#endif

//CWordFrequency.cpp
#include "CWordFrequency.h"

CWordFrequency::CWordFrequency() : countBeginning(0), countEnd(0),
countWithin(0)
{

}

CWordFrequency::~CWordFrequency()
{

}

void CWordFrequency::incrementCountBeginning()
{

    ++countBeginning;

}

void CWordFrequency::incrementCountEnd()
{

    ++countEnd;

}

void CWordFrequency::incrementCountWithin()
{

    ++countWithin;

}

int CWordFrequency::returnCountBeginning()
{

    return countBeginning;

}

int CWordFrequency::returnCountEnd()
{

    return countEnd;

}

int CWordFrequency::returnCountWithin()
{

    return countWithin;

}

//CRandomName.h
#include <fstream>
#include <map>
#include <string>
#include <cstdlib>
#include <ctime>
#include <vector>
#include <algorithm>

#include "CWordFrequency.h"

#ifndef CRANDOMNAME_H_
#define CRANDOMNAME_H_

class CRandomName
{

    private:

        std::string errorMessage;
        std::ifstream *fileStreamIn;
        std::ofstream *fileStreamOut;
        std::map<char, std::map<char, CWordFrequency> > baseMap;
        std::map<char, CWordFrequency> sequenceFrequencyMap;
        std::vector<char> startChars;
        std::vector<char> availableChars;
        CWordFrequency tempFrequency;

    public:

        CRandomName();
        ~CRandomName();
        void inputFile(std::ifstream &streamHandle);
        void processFile();
        void outputList(std::ofstream &streamHandle);
        std::string outputName(double minLength, double maxLength);

};

#endif

//CRandomName.cpp
#include <iostream>
#include <string>
#include <set>
#include <iterator>
#include "CRandomName.h"
using namespace std;

CRandomName::CRandomName()
{
 srand(time(NULL));
}

CRandomName::~CRandomName()
{
 fileStreamOut->close();
 fileStreamIn->close();
}

void CRandomName::inputFile(std::ifstream &streamHandle)
{

    fileStreamIn = &streamHandle;

}

//Set the boolean to true if you have a file that contains names with whitespaces. This does mean that the
//end of each name should be marked by a end of line character.
//Also the words will be converted to uppercase automatically
void CRandomName::processFile(bool NoWhiteSpaceSkip)
{

    std::string word;
    std::set<char> startCharsSet; //these sets are used to list the different starting and available chars.
    std::set<char> availableCharsSet; //they will be converted to vectors later.
                                       //I needed sets for their key ability, and vectors later because they
                                       //offer better access to the elements contained in the vector.
    char base;
    char sequence;
    int wordPosition;

    startChars.clear();
    availableChars.clear();

    while(fileStreamIn->good())
    {
       //Ignore whitespaces between words. Different names should be seperated by a '\n'
       if(NoWhiteSpaceSkip) {
          getline(*fileStreamIn,word);
       } else
           *fileStreamIn >> word;
       
        if(word.length() > 1) 
        {
           startCharsSet.insert(word[0]);
           availableCharsSet.insert(word[0]);
	        
           for (wordPosition = 0; (wordPosition + 1) < (word.length()); wordPosition++)
           {
              availableCharsSet.insert(word[wordPosition+1]);
              base = word[wordPosition];
              sequence = word[wordPosition + 1];

              CWordFrequency &wf = baseMap[base][sequence];
       
              if (wordPosition == 0) {wf.incrementCountBeginning();}
              else if ((wordPosition + 1) >= (word.length() - 1)) {wf.incrementCountEnd();}
              else if ((wordPosition > 0) && ((wordPosition + 1) < (word.length() - 1))) {wf.incrementCountWithin();}            
           }
        }
   }

   set<char>::iterator it;
   for ((it = startCharsSet.begin()); it != startCharsSet.end(); it++)
        startChars.push_back(*it);
   for ((it = availableCharsSet.begin()); it != availableCharsSet.end(); it++)
        availableChars.push_back(*it);
}

void CRandomName::outputList(std::ofstream &streamHandle)
{
   fileStreamOut = &streamHandle;

   std::map<char, std::map<char, CWordFrequency> >::iterator itr;

   std::map<char, CWordFrequency>::iterator itr2;

   for (itr = baseMap.begin(); itr != baseMap.end(); itr++)
   {
       sequenceFrequencyMap = itr->second;
       for (itr2 = sequenceFrequencyMap.begin(); itr2 != sequenceFrequencyMap.end(); itr2++)
       {
           tempFrequency = itr2->second;
           *fileStreamOut << itr->first << " " << itr2->first << " " 
               << tempFrequency.returnCountBeginning() << " " << 
               tempFrequency.returnCountWithin() << " " 
               << tempFrequency.returnCountEnd() << std::endl;
       }
   }
}

//Known bugs, if there are not enough names in the seed the program will crash.
std::string CRandomName::outputName(double minLength, double maxLength)
{
   std::string name;
   std::vector<char> freqVector;
   double range = static_cast<double>((maxLength - minLength) + 1);
   int rangeLength = static_cast<int>(minLength + (range * ((double)rand() / (double)(RAND_MAX + 1))));
   char a =  startChars.at(static_cast<int> (startChars.size() * rand() / ( RAND_MAX + 1.0 ) ));
   name += a;
   
   for(int counter = 1; counter < rangeLength; counter++)
   {
       int cdc = 0;
       if(baseMap.find(a) != baseMap.end())
       {
        for (int count = 0; count < availableChars.size(); count++)
        {
           char b = availableChars.at(count);
           if(baseMap[a].find(b) != baseMap[a].end())
           {
               if(counter == 1)
               {
                   for(int cc = 0; cc < (baseMap[a][b].returnCountBeginning()); cc++)
                   {
                       freqVector.push_back(b);
                       cdc++;
                   }
               }
               else if((counter + 1) >= (rangeLength - 1))
               {
                   for(int cc = 0; cc < baseMap[a][b].returnCountEnd();cc++)
                   {
                       freqVector.push_back(b);
                       cdc++;
                   }
               }
               else
               {
                   for(int cc = 0; cc < baseMap[a][b].returnCountWithin();cc++)
                   {
                       freqVector.push_back(b);
                       cdc++;
                   }
               }
           }
       }
   }
       std::random_shuffle(freqVector.begin(), freqVector.end());
       std::random_shuffle(freqVector.begin(), freqVector.end());
       std::random_shuffle(freqVector.begin(), freqVector.end());
       int c = (int)(((cdc) * rand() / ( RAND_MAX + 1.0 )));
       name += freqVector.at(c);
       a = freqVector.at(c);
   }
   return name;
}

//main.cpp
#include <iostream>
#include "CRandomName.h"

int main()
{
   CRandomName name;
   
   std::ifstream inFile("NameList.txt");
   std::ofstream outFile("Stats.txt", std::ios_base::trunc);
   
   name.inputFile(inFile);
   name.processFile();
   name.outputList(outFile);
   
   std::cout << name.outputName(3, 9) << '\n' << 
       name.outputName(3, 9) << '\n' << 
       name.outputName(3, 9) << '\n' << 
       name.outputName(3, 9) << '\n' << 
       name.outputName(3, 9) << '\n';
}

(The code is public domain)

And here is some sample output:

ONMY
TERLEMDA
ZENEEVML
CHIA
SAWH
SPTKIVCDE
ILAN
COONR
KETADAN
VIAN
XAIIN
URRIRERON
FRTOKRR
YONNAN
IDANSZEDS
CAALLSTF
FLVAHE
XHAM
DUSEXNCST
DEBAD
SHEUO
WARFA
MUASS
INEE

As you can see, some are ok but most aren't. This is because the dataset although big isn't quite big enough. To compensate for this, you could try to implement artificial rules to make the names more readable, but I will leave that as an exercise for the reader.

Markov-Chains Approach in Python

Here is the piece of code used in Mines of Elderlore to generate random names:

import random

# from http://www.geocities.com/anvrill/names/cc_goth.html
PLACES = ['Adara', 'Adena', 'Adrianne', 'Alarice', 'Alvita', 'Amara', 'Ambika', 'Antonia', 'Araceli', 'Balandria', 'Basha',
'Beryl', 'Bryn', 'Callia', 'Caryssa', 'Cassandra', 'Casondrah', 'Chatha', 'Ciara', 'Cynara', 'Cytheria', 'Dabria', 'Darcei',
'Deandra', 'Deirdre', 'Delores', 'Desdomna', 'Devi', 'Dominique', 'Drucilla', 'Duvessa', 'Ebony', 'Fantine', 'Fuscienne',
'Gabi', 'Gallia', 'Hanna', 'Hedda', 'Jerica', 'Jetta', 'Joby', 'Kacila', 'Kagami', 'Kala', 'Kallie', 'Keelia', 'Kerry',
'Kerry-Ann', 'Kimberly', 'Killian', 'Kory', 'Lilith', 'Lucretia', 'Lysha', 'Mercedes', 'Mia', 'Maura', 'Perdita', 'Quella',
'Riona', 'Safiya', 'Salina', 'Severin', 'Sidonia', 'Sirena', 'Solita', 'Tempest', 'Thea', 'Treva', 'Trista', 'Vala', 'Winta']

###############################################################################
# Markov Name model
# A random name generator, by Peter Corbett
# http://www.pick.ucam.org/~ptc24/mchain.html
# This script is hereby entered into the public domain
###############################################################################
class Mdict:
    def __init__(self):
        self.d = {}
    def __getitem__(self, key):
        if self.d.has_key(key):
            return self.d[key]
        else:
            raise KeyError(key)
    def add_key(self, prefix, suffix):
        if self.d.has_key(prefix):
            self.d[prefix].append(suffix)
        else:
            self.d[prefix] = [suffix]
    def get_suffix(self,prefix):
        l = self[prefix]
        return random.choice(l)  

class MName:
    """
    A name from a Markov chain
    """
    def __init__(self, chainlen = 2):
        """
        Building the dictionary
        """
        if chainlen > 10 or chainlen < 1:
            print "Chain length must be between 1 and 10, inclusive"
            sys.exit(0)
    
        self.mcd = Mdict()
        oldnames = []
        self.chainlen = chainlen
    
        for l in PLACES:
            l = l.strip()
            oldnames.append(l)
            s = " " * chainlen + l
            for n in range(0,len(l)):
                self.mcd.add_key(s[n:n+chainlen], s[n+chainlen])
            self.mcd.add_key(s[len(l):len(l)+chainlen], "\n")
    
    def New(self):
        """
        New name from the Markov chain
        """
        prefix = " " * self.chainlen
        name = ""
        suffix = ""
        while True:
            suffix = self.mcd.get_suffix(prefix)
            if suffix == "\n" or len(name) > 9:
                break
            else:
                name = name + suffix
                prefix = prefix[1:] + suffix
        return name.capitalize()  

for i in range(100):
    print MName().New()

Finite State Approach

source : Original author Jeff Lait

This generator uses a very simple finite state machine to select between individual letters in a fashion that creates pronounceable player names.

The comments at the top explain the nature of the beast. The only syllable based component is where I trigger the termination timer using a quickly accelerating random chance to keep names short.

Following code is in C++

#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>
#include <string.h>
#include <time.h>

//this function returns true if a random roll (1-100) is lower then c.
//So: rand_chance(20) has a 20% chance of returning true.
bool rand_chance(int c) {
	return( (rand() % 100 + 1) <= c);
}

//this function returns a random int between 0 and i-1.
int rand_choice(int i) {
	return rand() % i;
}

void rand_name(char *text, int len)
{
    // Very simple markov generator.
    // We repeat letters to make them more likely.
    const char *vowels = "aaaeeeiiiooouuyy'";
    const char *frictive = "rsfhvnmz";
    const char *plosive = "tpdgkbc";
    const char *weird = "qwjx";
    // State transitions..
    // v -> f, p, w, v'
    // v' -> f, p, w
    // f -> p', v
    // p -> v, f'
    // w, p', f' -> v

    int         syllables = 0;
    char        state;
    int         pos = 0;
    bool        prime = false;

    // Initial state choice
    if (rand_chance(30))
        state = 'v';
    else if (rand_chance(40))
        state = 'f';
    else if (rand_chance(70))
        state = 'p';
    else
        state = 'w';

    while (pos < len-1)
    {
        // Apply current state
        switch (state)
        {
            case 'v':
                text[pos++] = vowels[rand_choice(strlen(vowels))];
                if (!prime)
                    syllables++;
                break;
            case 'f':
                text[pos++] = frictive[rand_choice(strlen(frictive))];
                break;
            case 'p':
                text[pos++] = plosive[rand_choice(strlen(plosive))];
                break;
            case 'w':
                text[pos++] = weird[rand_choice(strlen(weird))];
                break;
        }

        // Chance to stop..
        if (syllables && pos >= 3)
        {
            if (rand_chance(20+pos*4))
                break;
        }

        // Transition...
        switch (state)
        {
            case 'v':
                if (!prime && rand_chance(10))
                {
                    state = 'v';
                    prime = true;
                    break;
                }
                else if (rand_chance(40))
                    state = 'f';
                else if (rand_chance(70))
                    state = 'p';
                else
                    state = 'w';
                prime = false;
                break;
            case 'f':
                if (!prime && rand_chance(50))
                {
                    prime = true;
                    state = 'p';
                    break;
                }
                state = 'v';
                prime = false;
                break;
            case 'p':
                if (!prime && rand_chance(10))
                {
                    prime = true;
                    state = 'f';
                    break;
                }
                state = 'v';
                prime = false;
                break;
            case 'w':
                state = 'v';
                prime = false;
                break;
        }
    }
    text[0] = toupper(text[0]);
    text[pos++] = '\0';

}

//Arguments:    First is max length of random name
//              Second is numer of random names
int main(int argc, char *argv[]) {
	srand ( time(NULL) );
	int j = 1;
	if(argc > 2) {
		j = atoi(argv[2]);
	} 
		
	while(j > 0) {
		if(argc > 1) {
			int i = atoi(argv[1]);
			char * a = (char*) malloc(sizeof(char) * i);
			rand_name(a, i);
			printf("%s\n",a);
		} else {
			char a [5];
			rand_name(a, 5);
			printf("%s\n",a);
		}
		j--;
	}
	return 0;
}

(The code is public domain)

This program accepts the following arguments, the first is the max length of the name, but most will be smaller. And the second is the number of names to generate. Defaults are 5 1.

Results:

./randomName 10 20
'tosezivc
Fdo
'dybze
Wask
Bof
Yqa
'vymi
Fatyope
Xar
Qow
Midu
Tezk
Tucik
Hgy
Eqyhp
Hti
Cevidi
Sd'gaquz
Ubeqy
Spa

As you can see the algorithm has a bias for short names, and is not always pronounceable. But it is a start. Also as this is a proof of concept implementation, the random number generator is not properly seeded. Running multiple types after each other could result in the same random names (depending on the operating system).

Syllable-based Approach

If you have one or more information about it, add it here.

Other Approaches